hubert-base-ls960-finetuned-gtzan

This model is a fine-tuned version of facebook/hubert-base-ls960 on the GTZAN dataset. It achieves the following results on the evaluation set:

• Loss: 0.6629
• Accuracy: 0.86

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 8
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 150
• num_epochs: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
4.0166	1.0	113	1.9303	0.37
2.9802	2.0	226	1.5313	0.5
2.7809	3.0	339	1.2783	0.58
1.6393	4.0	452	1.0212	0.71
1.7196	5.0	565	0.9676	0.7
1.7533	6.0	678	0.8288	0.72
1.2629	7.0	791	0.5269	0.85
0.9351	8.0	904	0.6159	0.82
0.5778	9.0	1017	0.5569	0.83
0.4505	10.0	1130	0.5305	0.86
0.5766	11.0	1243	0.7149	0.83
0.3385	12.0	1356	0.7123	0.86
0.1503	13.0	1469	0.7124	0.84
0.4639	14.0	1582	0.6589	0.84
0.0482	15.0	1695	0.6629	0.86

Framework versions

• Transformers 5.20.2
• Pytorch 2.11.0+cu130
• Datasets 4.8.5
• Tokenizers 0.23.3